The congestion control algorithms nowadays reached their limit of performance beyond today's requirements, so a new approach is required which can provide good utilisation in every environment. Upgrading the current TCP protocol was a good short term solution, but for long term, an entirely new system could be the answer. The point is to maintain equally optimal solution for different situations and different needs. According to a GENI recommendation, a protocol using forward error correction instead of congestion control can be more efficient and more robust on todays diversified networks.
In this thesis I will introduce you to a revolutionary new transport protocol based on erasure coding which is in development state. I will make comparison between the two systems, and the priority is the upgrading potential within. The new protocoll is being tested by a research team in the university. My task is to test the system in average situations, and to find the optimal environment for its best performance.
The point of this thesis is to summarize the major existing high speed TCP variants and compare their operation on specific parameters. The comparison uses the DFCP accomplished method in different situations.
At first I present the role and tasks of the transmission layer. I also demonstrate the most common congestion control procedures.I review the network simulation environment studying the above mentioned protocol as well.
To demonstrate the behavior of the protocol I model measurement cases. After evaluating these scenarios and taking the new experiences into consideration I deduct relevant consequences.
In the last chapter I determine eventuel innovations and future plans.